koft@elbereth.rutgers.edu (Dan Koft) (05/29/91)
Please help me if you can. I have a C program that I want to call an
IMSL routine called QDAGI, the code in FORTRAN looks like:
CALL QDAGI(F, BOUND, INTERV, ERRABS, ERRREL, RESULT, ERREST)
Where (from the IMSL manual):
F - User supplied function to be integrated. The form is F(X),
where: X - Independant variable (Input), F - The function
value (Output). F must be declared EXTERNAL in the calling
program.
BOUND - Finite bound of the integration range (Input). Ignored if
INTERV=2.
INTERV - Flag indicating integration range (Input). If INTERV is -1
the interval is -infinity to BOUND; if 1, BOUND to +infinity;
if 2 -infinity to +infinity.
ERRABS - Absolute accuracy desired (Input).
ERRREL - Relative accuracy desired (Input).
RESULT - Estimate of the integral from A to B of F (Output).
ERREST - Estimate of the absolute value of the error (Output).
I have tried all the ways I could think of to call this function
without proper results. Below is the test program from the IMSL
manual as I have translated it to C.
#include <stdio.h>
#include <math.h>
#define Abs(x) ( (x)>0? (x) : (-x))
double F(double x);
extern void qdagi(double *F(), double *bound, int *interv, double *errabs,
double *errel, double *result, double *errest);
main()
{
double abs,bound,errabs,errest,error,errrel,exact,ppi,result;
int interv = 1;
bound = 0.0;
errabs = 0.0;
errrel = 0.001;
qdagi((&)F(),&bound,&interv,&errabs,&errrel,&result,&errest);
/* The above line is where I believe the problem lies*/
ppi = 3.1415926;
exact = -ppi * log(10.)/20.;
error = Abs(result - exact);
printf(" Computed = %8.3f exact = %8.3f\nError ", result, exact);
printf("estimate = %f error = %f\n", errest, error);
}
double F(double x)
{
double temp;
temp = log(x)/(1. + pow((10. * x),2.0));
return temp;
}
--
[=============================================================================]
Dan Koft Email: koft@elbereth.rutgers.edu
RU-CS, Rutgers University {backbone}!rutgers!elbereth!koft
SERC, Busch Campus Phone: (908) 932-3216
Piscataway, NJ 08855
[=============================================================================]randolph@tuna.ssd.kodak.com (Gary L. Randolph) (05/30/91)
In article <May.29.12.00.35.1991.16903@elbereth.rutgers.edu> koft@elbereth.rutgers.edu (Dan Koft) writes: >Please help me if you can. I have a C program that I want to call an >IMSL routine called QDAGI, the code in FORTRAN looks like: > CALL QDAGI(F, BOUND, INTERV, ERRABS, ERRREL, RESULT, ERREST) ...stuff deleted... >I have tried all the ways I could think of to call this function >without proper results. Below is the test program from the IMSL >manual as I have translated it to C. >#include <stdio.h> >#include <math.h> >#define Abs(x) ( (x)>0? (x) : (-x)) >double F(double x); >extern void qdagi(double *F(), double *bound, int *interv, double *errabs, > double *errel, double *result, double *errest); **********************^^^^^^^^^^^^^^^^^^^^**************** First, your declaration of the formal argument, F, is incorrect. What you want is to declare F as pointer to function taking/returning double. extern void qdagi(double (*F)(double), ...rest as above) The parentheses are NOT optional! >main() >{ >double abs,bound,errabs,errest,error,errrel,exact,ppi,result; >int interv = 1; >bound = 0.0; >errabs = 0.0; >errrel = 0.001; >qdagi((&)F(),&bound,&interv,&errabs,&errrel,&result,&errest); >/* The above line is where I believe the problem lies*/ Yes, this is a problem. To pass the address of the function, you can pass either F or &F (the difference has recently been discussed here or .c++) So, the call would be: qdagi(F, ...same as before); To provide a simpler, yet silly, example: double f(double x){ printf("%s %lf", "\nThe double: ", x); return x;} void fun(double (*F)(double), double a){printf("%s %lf","\nThe DOUBLE: ",(*F)(a));} main(){ fun(f,55);} which results in: The double: 55.000000 The DOUBLE: 55.000000 Gary Randolph Eastman Kodak Company . . . . . . . .
khb@chiba.Eng.Sun.COM (Keith Bierman fpgroup) (05/30/91)
>calling IMSL/fortran library from C ....
The folks at IMSL have a C product as well nowadays. In addition, I
beleive that in at least one of their copious documents they deal with
the interlanguage calling issues a bit (it is platform specific). You
may find some material of interest in the Sun FORTRAN User's guide on
this topic as well.
--
----------------------------------------------------------------
Keith H. Bierman keith.bierman@Sun.COM| khb@chiba.Eng.Sun.COM
SMI 2550 Garcia 12-33 | (415 336 2648)
Mountain View, CA 94043